Thursday, 14. April 2011 14.04.11 14:09 Age: 7 Jahre

All data at one glimpse
The Department of Computer Science is researching the medical visualization of coronary heart diseases

Photo (Katharina Bätz): Stephan Arens in front of a graphic representation that enables the quick recognition of plaque in the coronary.

Paderborn. Coronary heart diseases are the most common cause of death in industrial nations. When the coronary vessels in the human body are plied with debris (“plaque”) the administration of oxygen to the heart is downgraded it can not only cause a shortness of breath but a cardiac death, too.

To obviate that such a condition comes about the professorship “Computer Graphics, Visualization and Image Processing“ of Prof. Dr. Gitta Domik at the Department of Computer Science at the University of Paderborn works on the improvement of early recognition of coronary heart diseases and the scheduling of operations. The challenge is the quick recognition and localization of plaque in the coronary.

“At stake is how to process the medical data so that the treating doctor is able to recognize from the pictures if there are dangerous constrictions in the coronary vessels or if they are building up”, explains the degreed Computer Scientist Stephan Arens, who is Scientific Assistant in the research group.

In doing so it has to be considered that not all plaques hold dangerous potential.  "Therefore it is exciting to be able to tell in the front end whether the plaque could get dangerous or not”, says Arens. “This decision cannot yet be foretold by any computer but only by an experienced doctor.”

The doctor wants to deepen his diagnosis to the specialties of the coronaries of his patient and wants to recognize the boundaries and exact location of the plaque. No superfluous information should delay him from this – and if lone for the reasons of time, as according to experience a diagnosis should be completed in ten minutes. This technical challenge, to find a time saving method for the accruing huge amounts of data of a breast-CT which enables a treating doctor to make a quick and fitting diagnose, is the task of the computer scientists.

The so called Curved Planar (CPR) is the most important technique of visualization for this purpose at the moment. It reduces the huge amount of data and filters just the relevant data out of the information-flood. The advanced form of CPR at the University of Paderborn enables a fade-out of all the non-relevant organs in the breast section and focuses on the coronary: the coronary arteries, which span the heart-muscle can be unrolled, hauled and turned around the axis, so that an exact localization of the plaques and the estimation of the distances between the plaque is going to be possible. The latter is also very important for the schedule of an operation for the insertion of stents and bypasses. Behind this algorithm is a lot of mathematic and clever programming of  fast graphic boards.

The performance of such a technique of visualization can just be measured with the achieved results of the medic. The “controlled User-experiments” which Stephan Arens has frequently accomplished in recent times, could help to make time and quality of diagnoses from the laboratory measurable. As they are very complex the  professorship Domik is amongst only a few who currently engage in this field. In spite of the fact, that the importance of quality control of visual display, especially in the field of medical diagnosis, is more and more strongly demanded.

In order to exchange medical data and know-how there is a cooperation between the heart- and diabetes-Centre NRW in Bad Oeynhausen since 1995. The the scientists of the institute of computer science therefore develop the software “Volume Studio”, which is exerted for researching purposes in the heart- and diabetes-Centre.


Contakt:
Prof. Dr. Gitta Domik 
Research Group Computer Graphics, Visualization and Image Processing
+49 (0)5251 60-6610
domik[at]uni-paderborn.de